On Tue, 2024-01-02 at 16:06 -0800, Andrii Nakryiko wrote: [...] > For the fix that Eduard proposed (and checking > try_match_pkt_pointers), should we do a similar simplification as we > do for scalar register comparisons? Make sure that data_end is always > on the right by swapping, if that's not the case. And also use > corresponding rev_opcode() and flip_opcode() operations to minimize > the amount of logic and duplicated code? > > And I mean not just for new JEQ/JNE cases, but let's also refactor and > simplify existing logic as well? Yes, this should simplify the function significantly.
